Pod is not respecting the memory limits set on the deployment.

In one of the deployment, the limits.memory is set to 512Mi, but the memory utilization by pod is 727.1Mi
~~~
$ oc adm top pods -n publix-s0eauths-dev
NAME                                       CPU(cores) MEMORY(bytes)
eauth-authorization-api-74c6c45d75-vzfb6   5m         727.1Mi
~~~

It is expected that the pod should not consume more memory than the limit. Also if the consumption is more then the pod should get terminated with OOM, but the pod is still in running state.

OCP 4.13.z

Pod is consuming more memory than the limit set in deployment.

Pod should not consume more memory than the limit set in deployment.

Below are the deployment yaml and pod yaml for reference :

Deployment yaml :
~~~
        resources:
          limits:
            cpu: 100m
            memory: 512Mi
          requests:
            cpu: 100m
            memory: 128Mi
~~~

Pod yaml :
~~~
    name: eauth-authorization-api
    ports:
    - containerPort: 8080
      name: web-api-port
      protocol: TCP
    resources:
      limits:
        cpu: 100m
        memory: 512Mi
      requests:
        cpu: 100m
        memory: 128Mi
~~~

Customer re-deployed the pod. Still the memory consumption is more than the limit :
~~~
$ oc adm top pods -n publix-s0eauths-dev
NAME                                     CPU(cores) MEMORY(bytes)
eauth-authorization-api-854ddbcdf6-p9w2t 5m         752Mi
~~~